Nick Lee

100K posts

Nick Lee

Nick Lee

@TeslaPLee

เข้าร่วม Eylül 2023
493 กำลังติดตาม1.6K ผู้ติดตาม
Nick Lee รีทวีตแล้ว
NVIDIA Robotics
NVIDIA Robotics@NVIDIARobotics·
Exciting news for Jetson developers 🎉 Gemma 4 is now on Jetson. @GoogleGemma’s latest multimodal, multilingual models run across the full Jetson platform—from Orin Nano to Thor—bringing on-device AI to robotics, edge, and embedded systems. Cut latency, manage costs, and keep sensitive data secure. Check out the tutorial and download the container to get started: jetson-ai-lab.com/models/
Google DeepMind@GoogleDeepMind

Meet Gemma 4: our new family of open models you can run on your own hardware. Built for advanced reasoning and agentic workflows, we’re releasing them under an Apache 2.0 license. Here’s what’s new 🧵

English
3
6
30
2.5K
Nick Lee รีทวีตแล้ว
Ilir Aliu
Ilir Aliu@IlirAliu_·
Your cleaner just got a robot coworker. And it’s already in real homes. @XSquareRobot just launched China’s first robot-assisted home cleaning service with 58[.]com. Finally some robot companies doing deployments! When you book a cleaning, a human shows up with a robot. They split the work. • human handles judgment-heavy tasks • robot handles repetitive work like wiping and tidying • both operate together in real environments This robot runs on an end-to-end foundation model. No fixed scripts. No pre-programmed paths. It sees, plans, and acts. Most robotics still lives in controlled environments. Like factories, labs or warehouses. Homes are different. They are messy, totally unstructured, you could say unpredictable. If a robot works there, it works almost anywhere. This is the shift from: robots that can work to robots that can live and operate alongside us. And it’s starting with something simple: cleaning your living room.
English
3
8
25
2.2K
Nick Lee รีทวีตแล้ว
Vala Afshar
Vala Afshar@ValaAfshar·
Nvidia CEO: “you cannot show me a task that is beneath me.” The enemy of continuous growth is arrogance, zero sum mindset and a sense of entitlement.
English
4
25
102
9.6K
Nick Lee รีทวีตแล้ว
Rapid Response 47
Rapid Response 47@RapidResponse47·
ZXX
284
711
5.7K
364.7K
Nick Lee รีทวีตแล้ว
Nick Lee รีทวีตแล้ว
Weizhuo(Ken) Wang
Weizhuo(Ken) Wang@KenWangWeizhuo·
A person walks around campus for 5 hours with cameras. That's it. That's the training data. The result? A humanoid robot that traverses unseen buildings, crowds, and glass walls — zero robot data, zero finetuning. EgoNav is here. egonav.weizhuowang.com None of these behaviors were pre-programmed: • Waiting for a door to open before entering • Steering around glass walls invisible to depth sensors • Yielding to pedestrians and resuming • Re-routing when furniture is rearranged All emerged from 5 hours of a human walking around. The prior is real. (1/6) #Humanoid #Robotics #DiffusionModel #EgoNav
English
11
50
239
26.7K
Nick Lee รีทวีตแล้ว
Willow
Willow@WillowVoiceAI·
Introducing Atlas 1. Willow's new frontier speech-to-text model. It outperforms ElevenLabs, Deepgram, OpenAI, and more by a wide margin. Built on the first scalable, human-powered transcription infrastructure ever built for real-time dictation.
English
233
145
2.2K
762K
Nick Lee รีทวีตแล้ว
Trade Whisperer
Trade Whisperer@TradexWhisperer·
NAND flash prices jumped 40% QoQ to an all-time high. 15 months straight. Suppliers shifting to high-layer 3D NAND is tightening legacy supply hard. SLC and MLC prices surged 20% and 50%, respectively in March. Wild 🔥 $MU $SNDK
Trade Whisperer tweet media
Trade Whisperer@TradexWhisperer

So what's up with Micron Technology $MU? Why did so many analysts lower the price targets but still rated it as Bullish? A bit contradicting isn't it? Well, here's the answer. The DRAM spot prices have been mostly flat. The 'recovery' didn't come as soon as they initially thought. They are humans too and got euphoric too early. Remember, many predicted the rates to happen earlier this year instead of last week. The rates have a lot of influence to cyclicals. There is also this thing called DRAM contract prices which is different from spot prices. Contracts as the name implies, is an agreement between Micron and the buyer. These contract prices have actually gone up 5-10% which is bullish for Micron for near term BUT.. the next wave of contracts might be neutral as spot prices has been mostly flat, a good indicator for DRAM demand. So why do companies make contracts? Two reasons: it's more predictable in their balance sheet which in turn is acts as a hedge against unexpected DRAM price fluctuations. My suspicion is that the contract prices got inflated early this year due to the whole AI and Rate Cut euphoria and companies rushed in to make DRAM contracts before it goes up 'too much' So while DRAM contracts may not be a problem for Q4 (Weds), it's an uncertainty for Q1/Q2 2025. The recovery is on its way but the rate cuts didn't happen until just a week ago. I don't know how fast the the benefits of rate cuts will materialize to DRAM prices. So what else can $MU do? Answer: SHIP a TON of HBM (Super Fast AI Memory) to $NVDA, $AMD and $AVGO. In Q3 Micron made about $100M in HBM sales. I am expecting about $200-300M in Q4 HBM sales and I am guessing $400-500M for Q1 2025. HBM is orders of magnitude more expensive than DRAM and its highly margin accreditive (it adds shit ton of earnings per share). They are expecting billions for 2025 and having more than $NVDA as its customer base will help to re-negotiate the prices to a more premium (10% more in 2025). Furthermore, over the last 2 quarters, Micron has been making huge improvements on HBM wafer yields (as all semi companies do over time) and this will add extra margin to their dollars. The market reaction will be all about guidance. The double beat is a requirement. It's already baked in. So IF Micron is able to surprise the market with a huge HBM sales guidance for the next a couple of quarters, market could react positively. I think elevated DRAM contract prices will not come until 2nd half of 2025. Neutral DRAM + Bullish HBM could move the needle. On our charts, $MU is still BLUE candled, meaning its bullish cycle remains intact. Remember playing earnings can be nerve-wracking even if you have a high conviction. I am long in $MU, we are not selling until DRAM prices have fully recovered.

English
3
10
152
18.6K
Nick Lee รีทวีตแล้ว
delian
delian@zebulgar·
TBPN HAS BEEN ACQUIRED BY OPENAI WHAT A TIME TO BE ALIVE GOOD WORK GENTLEMEN, WHAT A RUN
John Coogan@johncoogan

TBPN has been acquired by OpenAI! The show is staying the same and we’ll continue to go live at 11am pacific every weekday. This is a full circle moment for me as I’ve worked with @sama for well over a decade. He funded my first company in 2013. Then helped us fix a serious logjam during a critical funding round a few years later. When I took my second company through YC, he was president at the time, and then when I joined Founders Fund, the first deal I saw in motion was the post-ChatGPT round in late 2022. And as we started growing TBPN last year, he was the very first lab lead to join the show. Thank you to everyone that has been a part of TBPN until now. The last year has been the most fun and rewarding part of my career and we’re excited to have more resources than ever going forward.

English
12
7
651
89.8K
Nick Lee รีทวีตแล้ว
Ilir Aliu
Ilir Aliu@IlirAliu_·
Tesla/Figure: “We will have robots around us very soon…” 🚀 Some random Shenzhen startup:
English
7
26
112
9K
Nick Lee รีทวีตแล้ว
Trade Whisperer
Trade Whisperer@TradexWhisperer·
$TSM TSMC’s 3nm capacity is now extremely tight. Only “long-term loyal” customers get meaningful supply. Apple and Nvidia dominate allocation, while AMD, Intel and other ASIC makers get scraps. The 3nm club is invite-only.
Trade Whisperer tweet media
Trade Whisperer@TradexWhisperer

$TSM $NVDA $MU "TSMC accelerates fab construction with up to 10 facilities targeted for 2026" (Liberty Times Supply Chain sources) AI Party Just Started

English
7
4
69
11.5K
Nick Lee รีทวีตแล้ว
Ryan Petersen
Ryan Petersen@typesfast·
SpaceX is changing the world with 230 Mbps internet in the sky.
Ryan Petersen tweet media
English
11
3
255
10.9K
Nick Lee รีทวีตแล้ว
Luca Greco
Luca Greco@lucagrecoita·
3 major automakers are now deploying humanoid robots on production lines Here is why the die-casting industry should pay attention Xiaomi's robot just completed 3 consecutive hours of autonomous operation at a nut installation station in its EV Gigacasting workshop in 76 second cycle time, over 90% success rate. BMW deployed humanoid robots at Leipzig for high-voltage battery assembly. Toyota commercially deployed Agility Robotics' Digit at Woodstock for RAV4 production under a Robots as a Service model. Morgan Stanley forecasts the global humanoid robot market reaching $5 trillion by 2050. Why am I talking about humanoid robots? I'm no AI expert. I honestly have no idea whether the software will ever get good enough. But on one point I'm completely confident: if that software breakthrough happens, it means a multi-billion dollar opportunity for the die-casting industry in North America and Europe. Lightweight, high-precision structural parts in aluminum and magnesium, perfect for agile, affordable mass-manufactured robots. I want you positioned to capture as much of that upside as possible. 📰 Full article: industryarsenal.com/p/humanoid-rob… 📊 Gigacasting capacity is being built across China, Europe, North America, and Southeast Asia. The Gigacasting Database maps every machine, every facility, and every program: industryarsenal.com/database
English
1
7
21
2.1K
Nick Lee รีทวีตแล้ว
Mo Islam
Mo Islam@itsmoislam·
From our live show w/ @NYSE this week: Mike Annunziata (@nunzi46) of Also Capital (@CapitalAlso) on why he invested in K2 Space (@K2SpaceCo).
English
0
5
17
6.2K
Nick Lee รีทวีตแล้ว
a16z
a16z@a16z·
"The grid itself is civilization." Radiant CEO Doug Bernauer on productizing nuclear energy: "Nuclear reactors as products has never been seen before." "They're always usually these giant mega projects where you dig a huge hole in the ground and you take 5 to 10 years." "We're targeting one per week coming off of a production line." "Our product is for off the grid. So a megawatt reactor on a trailer." "We drive it or fly it to where the customer wants it to go and then turn it on within like 48 hours." "It lasts 5 years… 2 million gallon diesel equivalent." "So it's sort of an unbelievable thing where you can grow the grid or put a micro grid anywhere." @DougBernauer @RadiantNuclear
English
6
30
205
29.8K
Nick Lee รีทวีตแล้ว
Nick Lee รีทวีตแล้ว
Lukas Ziegler
Lukas Ziegler@lukas_m_ziegler·
Robots fixing cleaning robots! 🤯 Can we call this at this point a robotception? @GeneralistAI just released GEN-1, the first general-purpose robot AI model that truly masters physical tasks. The results are 99% success rates on tasks where previous models achieved 64%, completes tasks 3x faster than state of the art, and needs only 1 hour of training data per task. Real demonstrations at full speed, fully autonomous: folding t-shirts 86 times in a row, servicing robot vacuums 200+ times in a row, packing blocks 1,800 times in a row, folding boxes 200+ times in a row. Speed matters. GEN-1 folds boxes in 12 seconds versus 34 seconds for previous models. Phone packing takes 15 seconds versus 43 seconds. The robot completes tasks faster than human demonstrations. The training approach is different. GEN-1 learns from over 500,000 hours of real-world data, but not from robots. It learns from wearable devices recording humans doing millions of everyday activities. This transfers to robots with just 1 hour of robot-specific training per task. When something goes wrong, GEN-1 improvises. If a part gets bumped out of position, the robot figures out whether to regrasp it, use its other hand, or leverage the environment to recover. These creative solutions weren't explicitly programmed. Five months ago, GEN-0 proved that robot AI follows scaling laws like language models. GEN-1 proves the scaling works. Congrats @peteflorence, and team!👏🏼 ~~ ♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com
English
6
33
127
8.1K